CompSci 100E 15.1 What is a Binary Search?  Magic!  Has been used a the basis for “magical” tricks  Find telephone number ( without computer ) in seconds  If that isn’t magic, what is?  There are less than atoms in the universe.  If ordered, how long to locate a particular one?  Demo!  What is the big-Oh?

CompSci 100E 15.2 Analysis: Algorithms and Data Structures  We need a vocabulary to discuss performance and to reason about alternative algorithms and implementations  It’s faster! It’s more elegant! It’s safer! It’s cooler!  We need empirical tests and analytical/mathematical tools  Given two methods, which is better? Run them to check. o 30 seconds vs. 3 seconds, easy. 5 hours vs. 2 minutes, harder o What if it takes two weeks to implement the methods?  Use mathematics to analyze the algorithm,  The implementation is another matter, cache, compiler optimizations, OS, memory,…

CompSci 100E 15.3 Recursion and recurrences  Why are some functions written recursively?  Simpler to understand, but …  Mt. Everest reasons  Are there reasons to prefer iteration?  Better optimizer: programmer/scientist v. compiler  Six of one? Or serious differences o “One person’s meat is another person’s poison” o “To each his own”, “Chacun a son gout”, …  Complexity (big-Oh) for iterative and recursive functions  How to determine, estimate, intuit

CompSci 100E 15.4 What’s the complexity of this code? // first and last are integer indexes, list is List int lastIndex = first; Object pivot = list.get(first); for(int k=first+1; k <= last; k++){ Comparable ko = (Comparable) list.get(k); if (ko.compareTo(pivot) <= 0){ lastIndex++; Collections.swap(list,lastIndex,k); }  What is big-Oh cost of a loop that visits n elements of a vector?  Depends on loop body, if body O(1) then …  If body is O(n) then …  If body is O(k) for k in [0..n) then …

CompSci 100E 15.5 FastFinder.findHelper private Object findHelper(ArrayList list, int first, int last, int kindex){ int lastIndex = first; Object pivot = list.get(first); for(int k=first+1; k <= last; k++){ Comparable ko = (Comparable) list.get(k); if (ko.compareTo(pivot) <= 0){ lastIndex++; Collections.swap(list, lastIndex,k); } Collections.swap(list, lastIndex, first); if (lastIndex == kindex) return list.get(lastIndex); if (kindex < lastIndex) return findHelper(list, first, lastIndex-1, kindex); return findHelper(list, lastIndex+1, last, kindex); }

CompSci 100E 15.6 Different measures of complexity  Worst case  Gives a good upper-bound on behavior  Never get worse than this  Drawbacks?  Average case  What does average mean?  Averaged over all inputs? Assuming uniformly distributed random data?  Drawbacks?  Best case  Linear search, useful?

CompSci 100E 15.7 Multiplying and adding big-Oh  Suppose we do a linear search then we do another one  What is the complexity?  If we do 100 linear searches?  If we do n searches on a vector of size n?  What if we do binary search followed by linear search?  What are big-Oh complexities? Sum?  What about 50 binary searches? What about n searches?  What is the number of elements in the list (1,2,2,3,3,3)?  What about (1,2,2, …, n,n,…,n)?  How can we reason about this?

CompSci 100E 15.8 Helpful formulae  We always mean base 2 unless otherwise stated  What is log(1024)?  log(xy) log(x y ) log(2 n ) 2 (log n)  Sums (also, use sigma notation when possible)  … + 2 k = 2 k+1 – 1 =  … + n = n(n+1)/2 =  a + ar + ar 2 + … + ar n-1 = a(r n - 1)/(r-1)= log(x) + log(y) y log(x) nlog(2) = n 2 (log n) = n k  i=0 2i2i n  i=1 i n-1  i=0 ar i

CompSci 100E 15.9 Recursion Review  Recursive functions have two key attributes  There is a base case, sometimes called the exit case, which does not make a recursive call  All other cases make recursive call(s), the results of these calls are used to return a value when necessary o Ensure that every sequence of calls reaches base case o Some measure decreases/moves towards base case o “Measure” can be tricky, but usually it’s straightforward  Example: sequential search in an ArrayList  If first element is search key, done and return  Otherwise look in the “rest of the list”  How can we recurse on “rest of list”?

CompSci 100E Sequential search revisited  What is complexity of sequential search? Of code below? boolean search(ArrayList list, int first, Object target) { if (first >= list.size()) return false; else if (list.get(first).equals(target)) return true; else return search(list,first+1,target); }  Why are there three parameters? Same name good idea? boolean search(ArrayList list, Object target){ return search(list,0,target); }

CompSci 100E Why we study recurrences/complexity?  Tools to analyze algorithms  Machine-independent measuring methods  Familiarity with good data structures/algorithms  What is CS person: programmer, scientist, engineer? scientists build to learn, engineers learn to build  Mathematics is a notation that helps in thinking, discussion, programming

CompSci 100E Recurrences  Summing Numbers int sum(int n) { if (0 == n) return 0; else return n + sum(n-1); }  What is complexity? justification?  T(n) = time to compute sum for n T(n) = T(n-1) + 1 T(0) = 1  instead of 1, use O(1) for constant time  independent of n, the measure of problem size

CompSci 100E Solving recurrence relations  plug, simplify, reduce, guess, verify? T(n) = T(n-1) + 1 T(0) = 1 T(n) = T(n-k) + k find the pattern! Now, let k=n, then T(n) = T(0)+n = 1+n  get to base case, solve the recurrence: O(n) T(n-1) = T(n-1-1) + 1 T(n) = [T(n-2) + 1] + 1 = T(n-2)+2 T(n-2) = T(n-2-1) + 1 T(n) = [(T(n-3) + 1) + 1] + 1 = T(n-3)+3

CompSci 100E Complexity Practice  What is complexity of Build ? (what does it do?) ArrayList build(int n) { if (0 == n) return new ArrayList(); // empty ArrayList list = build(n-1); for(int k=0;k < n; k++){ list.add(new Integer(n)); } return list; }  Write an expression for T(n) and for T(0), solve.

CompSci 100E Recognizing Recurrences  Solve once, re-use in new contexts  T must be explicitly identified  n must be some measure of size of input/parameter o T(n) is the time for quicksort to run on an n-element vector T(n) = T(n/2) + O(1) binary search O( ) T(n) = T(n-1) + O(1) sequential search O( ) T(n) = 2T(n/2) + O(1) tree traversal O( ) T(n) = 2T(n/2) + O(n) quicksort O( ) T(n) = T(n-1) + O(n) selection sort O( )  Remember the algorithm, re-derive complexity n log n n log n n n2n2